Magnets create an invisible field that attracts certain objects. A common example is a refrigerator. More important to our daily life, magnets can also store data in computers. By using the direction of the magnetic field (for example, up or down), the microscopic magnetic field can store a small amount of memory as zero or one – computer language. A potential game changer for next-generation microelectronics.
Scientists at the US Department of Energy (DOE) Argonne National Laboratory want to replace bar magnets with small magnetic vortices. As small as three billionths of a meter, these vortices are called skyrmions, which form magnetic particles. One day, they may introduce a new generation of microelectronics for memory storage and high-performance computing.
“The bar magnet in computer memory is like a shoelace tied with a string; it takes almost as much energy to recycle them,” said Arthur McCray, a graduate student at Northwestern University who works in Argonne’s Materials Science Division (MSD). And every bar magnet doesn’t work due to chaos will affect others.
“In contrast, skyrmions are like two shoelaces tied together. No matter how tight you tighten the shoelaces, the shoelaces remain attached. Another important thing is that scientists can control their behavior by changing the temperature or applying electricity.
Scientists have a lot to learn about how skyrmions behave under different conditions. To study them, the Argonne-led team developed an artificial intelligence (AI) program that works with an advanced electron microscope at the Center for Nanoscale Materials (CNM), a DOE Office of Science facility at Argonne. Microscopes can visualize the atmosphere at very low temperatures.
The group’s magnets are a mixture of iron, germanium and tellurium. In its construction, this object resembles a stack of books with many pages. A collection of these sheets contains many skyrmions, and a single sheet can be cut up and analyzed at a facility like CNM.
“The CNM electron microscope combined with a type of AI called machine learning allowed us to visualize skyrmion particles and their behavior at different temperatures,” said Yue Li, a postdoctoral candidate at MSD.
“Our most exciting finding is that the metal is arranged in the expected order at minus 60 degrees Fahrenheit and above,” said Charudatta Phatak, a materials scientist and team leader at MSD. “But as we cool the sample, the structure of the skyrmions changes.” Like bubbles and beer foam, some skyrmions get bigger, some get smaller, some merge and some disappear.
At minus 270, the layer reached almost a perfect state of crisis, but the order returned when the temperature returned to minus 60. This makes a change in temperature changes that can be exploited in microelectronics in the future for the preservation of the memory.
“We estimate that the skyrmion energy efficiency could be 100 to 1000 times that of current memory in high-end computers used for research,” McCray said.
Energy efficiency is important for the next generation of microelectronics. Today’s microelectronics represents a large part of global energy consumption and could consume almost 25% in ten years. We need to find electronic devices that work well.
“We still have a long way to go before the skyrmion ends up in low-end computer storage in the future,” Phatak said. “However, this kind of innovative thinking about microelectronics is the key to the next generation of devices.”
This research was supported by the DOE Office of Basic Energy Sciences. The group’s machine learning programs are conducted on supercomputing facilities at the Argonne Leadership Computing Facility, a DOE science staff facility.
